Landing page and speed: where paid clicks get lost
Raising conversion rate by a tenth is the same as making every click a tenth cheaper — except it lasts and nobody can outbid you. What we actually know about the effect of speed, and which of it is recycled myth.
- Improving conversion rate works like a discount on every click, but it lasts and nobody competes for it in an auction.
- The effect of speed on bounce is well documented; the specific percentages from blogs contradict each other.
- Fix the match between ad and page first, then worry about milliseconds.
Money put into a budget buys more clicks. Money put into a landing page lowers the cost of every future conversion — and nobody can outbid you for it. At a 2 % conversion rate, improving to 2.2 % has the same effect as a ten per cent discount on every click, except permanently.
Campaigns still get fixed first. This article is about what to fix on the landing page and in what order — and about which numbers on speed to believe.
What we actually know about speed
This topic suffers from the same two data points being recycled for a decade and mutating along the way. What has support:
- Google, mobile load analysis. Extending load time from one second to ten raises the probability of a bounce by 123 %. From one to three seconds, by 32 %.
- Vodafone Italy (a web.dev case study): improving the LCP metric by 31 % on key pages led to 8 % higher sales from those pages.
- Deloitte and Google, Milliseconds Make Millions. The source of the widely quoted rule of thumb that one second costs roughly 7 % of conversions. Sample of 37 brands, 30.5 million sessions, measured hourly over 30 days.
Many “2026 statistics” about speed are recycled older data or contradict each other outright. For one frequently cited study we found two irreconcilable versions: conversion rate falling from 3.05 % to 0.41 % between the first and fifth second, and elsewhere “a 4.42 % drop per second of load time”. Both cannot be true.
The more important caveat: most of those studies are cross-sectional, comparing fast and slow sites. Fast sites also tend to be better built and better funded, so several things are being measured at once. Only controlled tests like Vodafone’s are causal.
The practical conclusion: speed has a demonstrable effect, especially on mobile and in the first seconds. Nobody can give you the exact percentage for your site — and more importantly, speed is not the first thing to fix.
The order of fixes by return
1. The match between ad and page
The most common and most expensive mistake: the ad promises “air conditioning service within 24 hours in Leeds” and the link goes to the homepage. The person has to hunt for what they were promised, and some of them will not.
The check: go through your ten most expensive keywords and look at where each one lands. If the page does not contain literally what the ad said, you have your first task — and it is free.
2. What the page says in the first three seconds
The visitor has to see immediately: what this is, who it is for, what it costs or starts from, and what to do next. The three most common failures:
- A headline that talks about you (“We are a traditional family firm”) instead of the customer.
- No price or at least a price range. For services this is the most common reason people leave.
- No visible action above the fold.
Copy has its own article: what belongs on a landing page.
3. Mobile
Most PPC clicks arrive from mobile and most pages are designed on a large monitor. Go through the purchase or enquiry on a phone, on a normal network, not on office wifi. Look for small buttons, forms with the wrong keyboard, the need to zoom, and full-screen pop-ups.
4. Speed
Only now. The usual culprits, in this order:
- Images. Unresized photographs and missing modern formats. The cheapest improvement there is.
- Third-party scripts. Chat, heatmaps, A/B tools, four tracking codes. Go through what is there and remove half.
- Custom fonts. Loaded without reserving space, so the page jumps as it loads.
- Pop-ups. A cookie banner and a newsletter offer at once, immediately on load.
5. Checkout or form
The most expensive point on the whole journey, because this is where people you have already paid for walk away. Forced registration, an unexpected delivery charge only at the last step, a missing preferred payment method. Go through it as a customer, all of it, on a phone.
How to measure whether it helped
- Test one thing. Change the headline, the image and the form together and you learn nothing.
- Work out how much data you need. At fifty conversions a month an A/B test is pointless; the difference you could detect would have to be enormous.
- Do not evaluate early. The first days tend to look most favourable and mean nothing.
- Measure on business. More enquiries of lower quality is not an improvement. For enquiry-driven sites, track how many turned into deals.
What to take away
The landing page is the one place in the whole system where an improvement faces no competition. Start with the match between ad and page, then clarity and mobile, and only at the end with milliseconds. And before you next raise a budget, walk the whole journey as a customer on a phone — it tends to be the cheapest audit there is.
Common questions
How much does site speed affect conversions?
The effect is documented, but nobody can give you a specific figure for your site in advance. Google’s published analysis of mobile load times reports that extending load from one to three seconds raises the probability of a bounce by 32 per cent, and from one to ten seconds by 123 per cent. The best causal evidence is the Vodafone Italy case study, where improving the LCP metric by 31 per cent led to 8 per cent higher sales. Beware frequently recycled percentages from blogs — many contradict each other, and most studies are cross-sectional, comparing fast and slow sites that differ in many other ways too.
What should I fix on a landing page first?
The match between the ad and the page. If the ad promises a specific service, timeframe or price and the link goes to a homepage or a generic category, you lose people before anything even loads. Go through your ten most expensive keywords and check that the destination page contains literally what the ad said. It is a free fix and usually has more impact than optimising speed.
Is A/B testing a landing page worth it?
Only when you have enough conversions. At fifty conversions a month you could reliably detect only enormous differences, so the test will usually conclude that nothing was proven — which gets misread as the change not working. At low volumes it is more useful to fix obvious faults, walk the customer journey on a phone, and measure on business rather than on conversion count. More enquiries of lower quality is not an improvement.
- Find Out How You Stack Up to New Industry Benchmarks for Mobile Page Speed, Think with GoogleSource of the figures on bounce probability rising 32 % (1 → 3 s) and 123 % (1 → 10 s). Google’s own analysis, that is a party with an interest in the topic of speed.
- Milliseconds Make Millions, Deloitte for GoogleSource of the “one second ≈ 7 % of conversions” rule of thumb. Sample of 37 brands and 30.5 million sessions over 30 days; results vary by sector and the study dates from 2019.
- Core Web Vitals Impact Your Online Business — An LCP Case Study, Blue TriangleContext on LCP improvement case studies including Vodafone Italy (LCP −31 %, sales +8 %). A vendor source; the original study is on web.dev.